Training in progress, epoch 2, checkpoint
Browse files- checkpoint-980/config.json +71 -0
- checkpoint-980/model.safetensors +3 -0
- checkpoint-980/optimizer.pt +3 -0
- checkpoint-980/preprocessor_config.json +14 -0
- checkpoint-980/rng_state.pth +3 -0
- checkpoint-980/scheduler.pt +3 -0
- checkpoint-980/trainer_state.json +1423 -0
- checkpoint-980/training_args.bin +3 -0
checkpoint-980/config.json
ADDED
|
@@ -0,0 +1,71 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_name_or_path": "vinai/PhoWhisper-small",
|
| 3 |
+
"activation_dropout": 0.0,
|
| 4 |
+
"activation_function": "gelu",
|
| 5 |
+
"apply_spec_augment": false,
|
| 6 |
+
"architectures": [
|
| 7 |
+
"ViSpeechClassification"
|
| 8 |
+
],
|
| 9 |
+
"attention_dropout": 0.0,
|
| 10 |
+
"begin_suppress_tokens": [
|
| 11 |
+
220,
|
| 12 |
+
50257
|
| 13 |
+
],
|
| 14 |
+
"bos_token_id": 50257,
|
| 15 |
+
"classifier_proj_size": 256,
|
| 16 |
+
"d_model": 768,
|
| 17 |
+
"decoder_attention_heads": 12,
|
| 18 |
+
"decoder_ffn_dim": 3072,
|
| 19 |
+
"decoder_layerdrop": 0.0,
|
| 20 |
+
"decoder_layers": 12,
|
| 21 |
+
"decoder_start_token_id": 50258,
|
| 22 |
+
"dropout": 0.0,
|
| 23 |
+
"encoder_attention_heads": 12,
|
| 24 |
+
"encoder_ffn_dim": 3072,
|
| 25 |
+
"encoder_layerdrop": 0.0,
|
| 26 |
+
"encoder_layers": 12,
|
| 27 |
+
"eos_token_id": 50257,
|
| 28 |
+
"forced_decoder_ids": null,
|
| 29 |
+
"id2label": {
|
| 30 |
+
"0": "male, northern dialect",
|
| 31 |
+
"1": "male, central dialect",
|
| 32 |
+
"2": "male, highland central dialect",
|
| 33 |
+
"3": "male, southern dialect",
|
| 34 |
+
"4": "female, northern dialect",
|
| 35 |
+
"5": "female, central dialect",
|
| 36 |
+
"6": "female, highland central dialect",
|
| 37 |
+
"7": "female, southern dialect"
|
| 38 |
+
},
|
| 39 |
+
"init_std": 0.02,
|
| 40 |
+
"is_encoder_decoder": true,
|
| 41 |
+
"label2id": {
|
| 42 |
+
"female, central dialect": "5",
|
| 43 |
+
"female, highland central dialect": "6",
|
| 44 |
+
"female, northern dialect": "4",
|
| 45 |
+
"female, southern dialect": "7",
|
| 46 |
+
"male, central dialect": "1",
|
| 47 |
+
"male, highland central dialect": "2",
|
| 48 |
+
"male, northern dialect": "0",
|
| 49 |
+
"male, southern dialect": "3"
|
| 50 |
+
},
|
| 51 |
+
"mask_feature_length": 10,
|
| 52 |
+
"mask_feature_min_masks": 0,
|
| 53 |
+
"mask_feature_prob": 0.0,
|
| 54 |
+
"mask_time_length": 10,
|
| 55 |
+
"mask_time_min_masks": 2,
|
| 56 |
+
"mask_time_prob": 0.05,
|
| 57 |
+
"max_length": 448,
|
| 58 |
+
"max_source_positions": 1500,
|
| 59 |
+
"max_target_positions": 448,
|
| 60 |
+
"median_filter_width": 7,
|
| 61 |
+
"model_type": "whisper",
|
| 62 |
+
"num_hidden_layers": 12,
|
| 63 |
+
"num_mel_bins": 80,
|
| 64 |
+
"pad_token_id": 50257,
|
| 65 |
+
"scale_embedding": false,
|
| 66 |
+
"torch_dtype": "float32",
|
| 67 |
+
"transformers_version": "4.41.2",
|
| 68 |
+
"use_cache": true,
|
| 69 |
+
"use_weighted_layer_sum": false,
|
| 70 |
+
"vocab_size": 51865
|
| 71 |
+
}
|
checkpoint-980/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:479d6a02b6e8ef1cc9bf55e7b450f1506879e462a5e882c949c9b0f858f4663f
|
| 3 |
+
size 358419776
|
checkpoint-980/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:d601a11859d70e2f3b6261aba308826146c8d805e33bc0878350084ba0db23e3
|
| 3 |
+
size 707744314
|
checkpoint-980/preprocessor_config.json
ADDED
|
@@ -0,0 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"chunk_length": 30,
|
| 3 |
+
"feature_extractor_type": "WhisperFeatureExtractor",
|
| 4 |
+
"feature_size": 80,
|
| 5 |
+
"hop_length": 160,
|
| 6 |
+
"n_fft": 400,
|
| 7 |
+
"n_samples": 480000,
|
| 8 |
+
"nb_max_frames": 3000,
|
| 9 |
+
"padding_side": "right",
|
| 10 |
+
"padding_value": 0.0,
|
| 11 |
+
"processor_class": "WhisperProcessor",
|
| 12 |
+
"return_attention_mask": false,
|
| 13 |
+
"sampling_rate": 16000
|
| 14 |
+
}
|
checkpoint-980/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:11dbe2e1d09a54af13a01e2c2fcce326cbf93290cb13f60ed7e2964460e4ca83
|
| 3 |
+
size 14244
|
checkpoint-980/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:2d93136dc2e1a872bc6972d91855a913393c3e1c6eb896f308953254a83fd1d3
|
| 3 |
+
size 1064
|
checkpoint-980/trainer_state.json
ADDED
|
@@ -0,0 +1,1423 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 0.8883299798792756,
|
| 3 |
+
"best_model_checkpoint": "PhoWhisper-small-vispeech-classifier-v4/checkpoint-490",
|
| 4 |
+
"epoch": 2.0,
|
| 5 |
+
"eval_steps": 500,
|
| 6 |
+
"global_step": 980,
|
| 7 |
+
"is_hyper_param_search": false,
|
| 8 |
+
"is_local_process_zero": true,
|
| 9 |
+
"is_world_process_zero": true,
|
| 10 |
+
"log_history": [
|
| 11 |
+
{
|
| 12 |
+
"epoch": 0.01020408163265306,
|
| 13 |
+
"grad_norm": 51640.66015625,
|
| 14 |
+
"learning_rate": 5.1020408163265303e-08,
|
| 15 |
+
"loss": 2.0778,
|
| 16 |
+
"step": 5
|
| 17 |
+
},
|
| 18 |
+
{
|
| 19 |
+
"epoch": 0.02040816326530612,
|
| 20 |
+
"grad_norm": 46773.35546875,
|
| 21 |
+
"learning_rate": 1.0204081632653061e-07,
|
| 22 |
+
"loss": 2.0809,
|
| 23 |
+
"step": 10
|
| 24 |
+
},
|
| 25 |
+
{
|
| 26 |
+
"epoch": 0.030612244897959183,
|
| 27 |
+
"grad_norm": 45343.12109375,
|
| 28 |
+
"learning_rate": 1.5306122448979592e-07,
|
| 29 |
+
"loss": 2.0771,
|
| 30 |
+
"step": 15
|
| 31 |
+
},
|
| 32 |
+
{
|
| 33 |
+
"epoch": 0.04081632653061224,
|
| 34 |
+
"grad_norm": 49285.91796875,
|
| 35 |
+
"learning_rate": 2.0408163265306121e-07,
|
| 36 |
+
"loss": 2.077,
|
| 37 |
+
"step": 20
|
| 38 |
+
},
|
| 39 |
+
{
|
| 40 |
+
"epoch": 0.05102040816326531,
|
| 41 |
+
"grad_norm": 63271.82421875,
|
| 42 |
+
"learning_rate": 2.5510204081632656e-07,
|
| 43 |
+
"loss": 2.0755,
|
| 44 |
+
"step": 25
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.061224489795918366,
|
| 48 |
+
"grad_norm": 51649.79296875,
|
| 49 |
+
"learning_rate": 3.0612244897959183e-07,
|
| 50 |
+
"loss": 2.0764,
|
| 51 |
+
"step": 30
|
| 52 |
+
},
|
| 53 |
+
{
|
| 54 |
+
"epoch": 0.07142857142857142,
|
| 55 |
+
"grad_norm": 51633.98828125,
|
| 56 |
+
"learning_rate": 3.5714285714285716e-07,
|
| 57 |
+
"loss": 2.0763,
|
| 58 |
+
"step": 35
|
| 59 |
+
},
|
| 60 |
+
{
|
| 61 |
+
"epoch": 0.08163265306122448,
|
| 62 |
+
"grad_norm": 61777.23828125,
|
| 63 |
+
"learning_rate": 4.0816326530612243e-07,
|
| 64 |
+
"loss": 2.0772,
|
| 65 |
+
"step": 40
|
| 66 |
+
},
|
| 67 |
+
{
|
| 68 |
+
"epoch": 0.09183673469387756,
|
| 69 |
+
"grad_norm": 58467.15625,
|
| 70 |
+
"learning_rate": 4.591836734693878e-07,
|
| 71 |
+
"loss": 2.0743,
|
| 72 |
+
"step": 45
|
| 73 |
+
},
|
| 74 |
+
{
|
| 75 |
+
"epoch": 0.10204081632653061,
|
| 76 |
+
"grad_norm": 48974.37890625,
|
| 77 |
+
"learning_rate": 5.102040816326531e-07,
|
| 78 |
+
"loss": 2.072,
|
| 79 |
+
"step": 50
|
| 80 |
+
},
|
| 81 |
+
{
|
| 82 |
+
"epoch": 0.11224489795918367,
|
| 83 |
+
"grad_norm": 71007.515625,
|
| 84 |
+
"learning_rate": 5.612244897959184e-07,
|
| 85 |
+
"loss": 2.0751,
|
| 86 |
+
"step": 55
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.12244897959183673,
|
| 90 |
+
"grad_norm": 58237.77734375,
|
| 91 |
+
"learning_rate": 6.122448979591837e-07,
|
| 92 |
+
"loss": 2.0702,
|
| 93 |
+
"step": 60
|
| 94 |
+
},
|
| 95 |
+
{
|
| 96 |
+
"epoch": 0.1326530612244898,
|
| 97 |
+
"grad_norm": 58785.83203125,
|
| 98 |
+
"learning_rate": 6.632653061224491e-07,
|
| 99 |
+
"loss": 2.0713,
|
| 100 |
+
"step": 65
|
| 101 |
+
},
|
| 102 |
+
{
|
| 103 |
+
"epoch": 0.14285714285714285,
|
| 104 |
+
"grad_norm": 52771.265625,
|
| 105 |
+
"learning_rate": 7.142857142857143e-07,
|
| 106 |
+
"loss": 2.069,
|
| 107 |
+
"step": 70
|
| 108 |
+
},
|
| 109 |
+
{
|
| 110 |
+
"epoch": 0.15306122448979592,
|
| 111 |
+
"grad_norm": 69647.078125,
|
| 112 |
+
"learning_rate": 7.653061224489796e-07,
|
| 113 |
+
"loss": 2.0684,
|
| 114 |
+
"step": 75
|
| 115 |
+
},
|
| 116 |
+
{
|
| 117 |
+
"epoch": 0.16326530612244897,
|
| 118 |
+
"grad_norm": 59148.47265625,
|
| 119 |
+
"learning_rate": 8.163265306122449e-07,
|
| 120 |
+
"loss": 2.0634,
|
| 121 |
+
"step": 80
|
| 122 |
+
},
|
| 123 |
+
{
|
| 124 |
+
"epoch": 0.17346938775510204,
|
| 125 |
+
"grad_norm": 66723.921875,
|
| 126 |
+
"learning_rate": 8.673469387755103e-07,
|
| 127 |
+
"loss": 2.064,
|
| 128 |
+
"step": 85
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.1836734693877551,
|
| 132 |
+
"grad_norm": 72545.203125,
|
| 133 |
+
"learning_rate": 9.183673469387756e-07,
|
| 134 |
+
"loss": 2.0593,
|
| 135 |
+
"step": 90
|
| 136 |
+
},
|
| 137 |
+
{
|
| 138 |
+
"epoch": 0.19387755102040816,
|
| 139 |
+
"grad_norm": 59112.55859375,
|
| 140 |
+
"learning_rate": 9.69387755102041e-07,
|
| 141 |
+
"loss": 2.0553,
|
| 142 |
+
"step": 95
|
| 143 |
+
},
|
| 144 |
+
{
|
| 145 |
+
"epoch": 0.20408163265306123,
|
| 146 |
+
"grad_norm": 68369.265625,
|
| 147 |
+
"learning_rate": 1.0204081632653063e-06,
|
| 148 |
+
"loss": 2.0566,
|
| 149 |
+
"step": 100
|
| 150 |
+
},
|
| 151 |
+
{
|
| 152 |
+
"epoch": 0.21428571428571427,
|
| 153 |
+
"grad_norm": 51080.62890625,
|
| 154 |
+
"learning_rate": 1.0714285714285714e-06,
|
| 155 |
+
"loss": 2.0505,
|
| 156 |
+
"step": 105
|
| 157 |
+
},
|
| 158 |
+
{
|
| 159 |
+
"epoch": 0.22448979591836735,
|
| 160 |
+
"grad_norm": 78760.0859375,
|
| 161 |
+
"learning_rate": 1.122448979591837e-06,
|
| 162 |
+
"loss": 2.0496,
|
| 163 |
+
"step": 110
|
| 164 |
+
},
|
| 165 |
+
{
|
| 166 |
+
"epoch": 0.23469387755102042,
|
| 167 |
+
"grad_norm": 57198.7890625,
|
| 168 |
+
"learning_rate": 1.1734693877551022e-06,
|
| 169 |
+
"loss": 2.0458,
|
| 170 |
+
"step": 115
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.24489795918367346,
|
| 174 |
+
"grad_norm": 98369.4609375,
|
| 175 |
+
"learning_rate": 1.2244897959183673e-06,
|
| 176 |
+
"loss": 2.0336,
|
| 177 |
+
"step": 120
|
| 178 |
+
},
|
| 179 |
+
{
|
| 180 |
+
"epoch": 0.25510204081632654,
|
| 181 |
+
"grad_norm": 80801.5703125,
|
| 182 |
+
"learning_rate": 1.2755102040816329e-06,
|
| 183 |
+
"loss": 2.0354,
|
| 184 |
+
"step": 125
|
| 185 |
+
},
|
| 186 |
+
{
|
| 187 |
+
"epoch": 0.2653061224489796,
|
| 188 |
+
"grad_norm": 82179.9765625,
|
| 189 |
+
"learning_rate": 1.3265306122448982e-06,
|
| 190 |
+
"loss": 2.0263,
|
| 191 |
+
"step": 130
|
| 192 |
+
},
|
| 193 |
+
{
|
| 194 |
+
"epoch": 0.2755102040816326,
|
| 195 |
+
"grad_norm": 83225.078125,
|
| 196 |
+
"learning_rate": 1.3775510204081633e-06,
|
| 197 |
+
"loss": 2.0168,
|
| 198 |
+
"step": 135
|
| 199 |
+
},
|
| 200 |
+
{
|
| 201 |
+
"epoch": 0.2857142857142857,
|
| 202 |
+
"grad_norm": 62451.64453125,
|
| 203 |
+
"learning_rate": 1.4285714285714286e-06,
|
| 204 |
+
"loss": 2.023,
|
| 205 |
+
"step": 140
|
| 206 |
+
},
|
| 207 |
+
{
|
| 208 |
+
"epoch": 0.29591836734693877,
|
| 209 |
+
"grad_norm": 75567.9453125,
|
| 210 |
+
"learning_rate": 1.479591836734694e-06,
|
| 211 |
+
"loss": 2.0206,
|
| 212 |
+
"step": 145
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.30612244897959184,
|
| 216 |
+
"grad_norm": 98461.484375,
|
| 217 |
+
"learning_rate": 1.5306122448979593e-06,
|
| 218 |
+
"loss": 1.9984,
|
| 219 |
+
"step": 150
|
| 220 |
+
},
|
| 221 |
+
{
|
| 222 |
+
"epoch": 0.3163265306122449,
|
| 223 |
+
"grad_norm": 75078.1484375,
|
| 224 |
+
"learning_rate": 1.5816326530612248e-06,
|
| 225 |
+
"loss": 1.9899,
|
| 226 |
+
"step": 155
|
| 227 |
+
},
|
| 228 |
+
{
|
| 229 |
+
"epoch": 0.32653061224489793,
|
| 230 |
+
"grad_norm": 97895.671875,
|
| 231 |
+
"learning_rate": 1.6326530612244897e-06,
|
| 232 |
+
"loss": 1.9758,
|
| 233 |
+
"step": 160
|
| 234 |
+
},
|
| 235 |
+
{
|
| 236 |
+
"epoch": 0.336734693877551,
|
| 237 |
+
"grad_norm": 65880.2421875,
|
| 238 |
+
"learning_rate": 1.6836734693877552e-06,
|
| 239 |
+
"loss": 1.9918,
|
| 240 |
+
"step": 165
|
| 241 |
+
},
|
| 242 |
+
{
|
| 243 |
+
"epoch": 0.3469387755102041,
|
| 244 |
+
"grad_norm": 70675.578125,
|
| 245 |
+
"learning_rate": 1.7346938775510206e-06,
|
| 246 |
+
"loss": 1.9715,
|
| 247 |
+
"step": 170
|
| 248 |
+
},
|
| 249 |
+
{
|
| 250 |
+
"epoch": 0.35714285714285715,
|
| 251 |
+
"grad_norm": 64327.19921875,
|
| 252 |
+
"learning_rate": 1.7857142857142859e-06,
|
| 253 |
+
"loss": 1.9597,
|
| 254 |
+
"step": 175
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.3673469387755102,
|
| 258 |
+
"grad_norm": 80610.6171875,
|
| 259 |
+
"learning_rate": 1.8367346938775512e-06,
|
| 260 |
+
"loss": 1.9632,
|
| 261 |
+
"step": 180
|
| 262 |
+
},
|
| 263 |
+
{
|
| 264 |
+
"epoch": 0.37755102040816324,
|
| 265 |
+
"grad_norm": 110899.375,
|
| 266 |
+
"learning_rate": 1.8877551020408163e-06,
|
| 267 |
+
"loss": 1.9476,
|
| 268 |
+
"step": 185
|
| 269 |
+
},
|
| 270 |
+
{
|
| 271 |
+
"epoch": 0.3877551020408163,
|
| 272 |
+
"grad_norm": 111132.5703125,
|
| 273 |
+
"learning_rate": 1.938775510204082e-06,
|
| 274 |
+
"loss": 1.9369,
|
| 275 |
+
"step": 190
|
| 276 |
+
},
|
| 277 |
+
{
|
| 278 |
+
"epoch": 0.3979591836734694,
|
| 279 |
+
"grad_norm": 88722.625,
|
| 280 |
+
"learning_rate": 1.989795918367347e-06,
|
| 281 |
+
"loss": 1.9452,
|
| 282 |
+
"step": 195
|
| 283 |
+
},
|
| 284 |
+
{
|
| 285 |
+
"epoch": 0.40816326530612246,
|
| 286 |
+
"grad_norm": 105874.296875,
|
| 287 |
+
"learning_rate": 2.0408163265306125e-06,
|
| 288 |
+
"loss": 1.9459,
|
| 289 |
+
"step": 200
|
| 290 |
+
},
|
| 291 |
+
{
|
| 292 |
+
"epoch": 0.41836734693877553,
|
| 293 |
+
"grad_norm": 97286.0078125,
|
| 294 |
+
"learning_rate": 2.0918367346938776e-06,
|
| 295 |
+
"loss": 1.918,
|
| 296 |
+
"step": 205
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.42857142857142855,
|
| 300 |
+
"grad_norm": 82848.0625,
|
| 301 |
+
"learning_rate": 2.1428571428571427e-06,
|
| 302 |
+
"loss": 1.9135,
|
| 303 |
+
"step": 210
|
| 304 |
+
},
|
| 305 |
+
{
|
| 306 |
+
"epoch": 0.4387755102040816,
|
| 307 |
+
"grad_norm": 114854.4609375,
|
| 308 |
+
"learning_rate": 2.1938775510204083e-06,
|
| 309 |
+
"loss": 1.8869,
|
| 310 |
+
"step": 215
|
| 311 |
+
},
|
| 312 |
+
{
|
| 313 |
+
"epoch": 0.4489795918367347,
|
| 314 |
+
"grad_norm": 86661.25,
|
| 315 |
+
"learning_rate": 2.244897959183674e-06,
|
| 316 |
+
"loss": 1.8861,
|
| 317 |
+
"step": 220
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.45918367346938777,
|
| 321 |
+
"grad_norm": 127591.3984375,
|
| 322 |
+
"learning_rate": 2.295918367346939e-06,
|
| 323 |
+
"loss": 1.8396,
|
| 324 |
+
"step": 225
|
| 325 |
+
},
|
| 326 |
+
{
|
| 327 |
+
"epoch": 0.46938775510204084,
|
| 328 |
+
"grad_norm": 71038.8671875,
|
| 329 |
+
"learning_rate": 2.3469387755102044e-06,
|
| 330 |
+
"loss": 1.8995,
|
| 331 |
+
"step": 230
|
| 332 |
+
},
|
| 333 |
+
{
|
| 334 |
+
"epoch": 0.47959183673469385,
|
| 335 |
+
"grad_norm": 101402.421875,
|
| 336 |
+
"learning_rate": 2.3979591836734696e-06,
|
| 337 |
+
"loss": 1.8699,
|
| 338 |
+
"step": 235
|
| 339 |
+
},
|
| 340 |
+
{
|
| 341 |
+
"epoch": 0.4897959183673469,
|
| 342 |
+
"grad_norm": 113157.3828125,
|
| 343 |
+
"learning_rate": 2.4489795918367347e-06,
|
| 344 |
+
"loss": 1.8599,
|
| 345 |
+
"step": 240
|
| 346 |
+
},
|
| 347 |
+
{
|
| 348 |
+
"epoch": 0.5,
|
| 349 |
+
"grad_norm": 115485.0546875,
|
| 350 |
+
"learning_rate": 2.5e-06,
|
| 351 |
+
"loss": 1.8214,
|
| 352 |
+
"step": 245
|
| 353 |
+
},
|
| 354 |
+
{
|
| 355 |
+
"epoch": 0.5102040816326531,
|
| 356 |
+
"grad_norm": 139369.8125,
|
| 357 |
+
"learning_rate": 2.5510204081632657e-06,
|
| 358 |
+
"loss": 1.8539,
|
| 359 |
+
"step": 250
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.5204081632653061,
|
| 363 |
+
"grad_norm": 141401.46875,
|
| 364 |
+
"learning_rate": 2.602040816326531e-06,
|
| 365 |
+
"loss": 1.7507,
|
| 366 |
+
"step": 255
|
| 367 |
+
},
|
| 368 |
+
{
|
| 369 |
+
"epoch": 0.5306122448979592,
|
| 370 |
+
"grad_norm": 125767.3984375,
|
| 371 |
+
"learning_rate": 2.6530612244897964e-06,
|
| 372 |
+
"loss": 1.7948,
|
| 373 |
+
"step": 260
|
| 374 |
+
},
|
| 375 |
+
{
|
| 376 |
+
"epoch": 0.5408163265306123,
|
| 377 |
+
"grad_norm": 148848.5625,
|
| 378 |
+
"learning_rate": 2.7040816326530615e-06,
|
| 379 |
+
"loss": 1.7871,
|
| 380 |
+
"step": 265
|
| 381 |
+
},
|
| 382 |
+
{
|
| 383 |
+
"epoch": 0.5510204081632653,
|
| 384 |
+
"grad_norm": 152509.4375,
|
| 385 |
+
"learning_rate": 2.7551020408163266e-06,
|
| 386 |
+
"loss": 1.8054,
|
| 387 |
+
"step": 270
|
| 388 |
+
},
|
| 389 |
+
{
|
| 390 |
+
"epoch": 0.5612244897959183,
|
| 391 |
+
"grad_norm": 114681.078125,
|
| 392 |
+
"learning_rate": 2.8061224489795917e-06,
|
| 393 |
+
"loss": 1.7711,
|
| 394 |
+
"step": 275
|
| 395 |
+
},
|
| 396 |
+
{
|
| 397 |
+
"epoch": 0.5714285714285714,
|
| 398 |
+
"grad_norm": 119104.984375,
|
| 399 |
+
"learning_rate": 2.8571428571428573e-06,
|
| 400 |
+
"loss": 1.7139,
|
| 401 |
+
"step": 280
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.5816326530612245,
|
| 405 |
+
"grad_norm": 123160.5,
|
| 406 |
+
"learning_rate": 2.908163265306123e-06,
|
| 407 |
+
"loss": 1.703,
|
| 408 |
+
"step": 285
|
| 409 |
+
},
|
| 410 |
+
{
|
| 411 |
+
"epoch": 0.5918367346938775,
|
| 412 |
+
"grad_norm": 157218.203125,
|
| 413 |
+
"learning_rate": 2.959183673469388e-06,
|
| 414 |
+
"loss": 1.7685,
|
| 415 |
+
"step": 290
|
| 416 |
+
},
|
| 417 |
+
{
|
| 418 |
+
"epoch": 0.6020408163265306,
|
| 419 |
+
"grad_norm": 134804.828125,
|
| 420 |
+
"learning_rate": 3.0102040816326534e-06,
|
| 421 |
+
"loss": 1.7053,
|
| 422 |
+
"step": 295
|
| 423 |
+
},
|
| 424 |
+
{
|
| 425 |
+
"epoch": 0.6122448979591837,
|
| 426 |
+
"grad_norm": 153977.625,
|
| 427 |
+
"learning_rate": 3.0612244897959185e-06,
|
| 428 |
+
"loss": 1.6711,
|
| 429 |
+
"step": 300
|
| 430 |
+
},
|
| 431 |
+
{
|
| 432 |
+
"epoch": 0.6224489795918368,
|
| 433 |
+
"grad_norm": 204690.125,
|
| 434 |
+
"learning_rate": 3.112244897959184e-06,
|
| 435 |
+
"loss": 1.5873,
|
| 436 |
+
"step": 305
|
| 437 |
+
},
|
| 438 |
+
{
|
| 439 |
+
"epoch": 0.6326530612244898,
|
| 440 |
+
"grad_norm": 158604.1875,
|
| 441 |
+
"learning_rate": 3.1632653061224496e-06,
|
| 442 |
+
"loss": 1.6642,
|
| 443 |
+
"step": 310
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.6428571428571429,
|
| 447 |
+
"grad_norm": 181877.484375,
|
| 448 |
+
"learning_rate": 3.2142857142857147e-06,
|
| 449 |
+
"loss": 1.6204,
|
| 450 |
+
"step": 315
|
| 451 |
+
},
|
| 452 |
+
{
|
| 453 |
+
"epoch": 0.6530612244897959,
|
| 454 |
+
"grad_norm": 157735.734375,
|
| 455 |
+
"learning_rate": 3.2653061224489794e-06,
|
| 456 |
+
"loss": 1.6152,
|
| 457 |
+
"step": 320
|
| 458 |
+
},
|
| 459 |
+
{
|
| 460 |
+
"epoch": 0.6632653061224489,
|
| 461 |
+
"grad_norm": 121935.3828125,
|
| 462 |
+
"learning_rate": 3.316326530612245e-06,
|
| 463 |
+
"loss": 1.6357,
|
| 464 |
+
"step": 325
|
| 465 |
+
},
|
| 466 |
+
{
|
| 467 |
+
"epoch": 0.673469387755102,
|
| 468 |
+
"grad_norm": 133579.859375,
|
| 469 |
+
"learning_rate": 3.3673469387755105e-06,
|
| 470 |
+
"loss": 1.6085,
|
| 471 |
+
"step": 330
|
| 472 |
+
},
|
| 473 |
+
{
|
| 474 |
+
"epoch": 0.6836734693877551,
|
| 475 |
+
"grad_norm": 154886.953125,
|
| 476 |
+
"learning_rate": 3.4183673469387756e-06,
|
| 477 |
+
"loss": 1.5609,
|
| 478 |
+
"step": 335
|
| 479 |
+
},
|
| 480 |
+
{
|
| 481 |
+
"epoch": 0.6938775510204082,
|
| 482 |
+
"grad_norm": 151655.0625,
|
| 483 |
+
"learning_rate": 3.469387755102041e-06,
|
| 484 |
+
"loss": 1.5905,
|
| 485 |
+
"step": 340
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.7040816326530612,
|
| 489 |
+
"grad_norm": 198471.609375,
|
| 490 |
+
"learning_rate": 3.5204081632653062e-06,
|
| 491 |
+
"loss": 1.5052,
|
| 492 |
+
"step": 345
|
| 493 |
+
},
|
| 494 |
+
{
|
| 495 |
+
"epoch": 0.7142857142857143,
|
| 496 |
+
"grad_norm": 191516.453125,
|
| 497 |
+
"learning_rate": 3.5714285714285718e-06,
|
| 498 |
+
"loss": 1.5798,
|
| 499 |
+
"step": 350
|
| 500 |
+
},
|
| 501 |
+
{
|
| 502 |
+
"epoch": 0.7244897959183674,
|
| 503 |
+
"grad_norm": 207511.296875,
|
| 504 |
+
"learning_rate": 3.6224489795918373e-06,
|
| 505 |
+
"loss": 1.5382,
|
| 506 |
+
"step": 355
|
| 507 |
+
},
|
| 508 |
+
{
|
| 509 |
+
"epoch": 0.7346938775510204,
|
| 510 |
+
"grad_norm": 170820.609375,
|
| 511 |
+
"learning_rate": 3.6734693877551024e-06,
|
| 512 |
+
"loss": 1.485,
|
| 513 |
+
"step": 360
|
| 514 |
+
},
|
| 515 |
+
{
|
| 516 |
+
"epoch": 0.7448979591836735,
|
| 517 |
+
"grad_norm": 183925.796875,
|
| 518 |
+
"learning_rate": 3.724489795918368e-06,
|
| 519 |
+
"loss": 1.427,
|
| 520 |
+
"step": 365
|
| 521 |
+
},
|
| 522 |
+
{
|
| 523 |
+
"epoch": 0.7551020408163265,
|
| 524 |
+
"grad_norm": 302829.75,
|
| 525 |
+
"learning_rate": 3.7755102040816327e-06,
|
| 526 |
+
"loss": 1.4242,
|
| 527 |
+
"step": 370
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.7653061224489796,
|
| 531 |
+
"grad_norm": 292990.5625,
|
| 532 |
+
"learning_rate": 3.826530612244898e-06,
|
| 533 |
+
"loss": 1.484,
|
| 534 |
+
"step": 375
|
| 535 |
+
},
|
| 536 |
+
{
|
| 537 |
+
"epoch": 0.7755102040816326,
|
| 538 |
+
"grad_norm": 189517.421875,
|
| 539 |
+
"learning_rate": 3.877551020408164e-06,
|
| 540 |
+
"loss": 1.5039,
|
| 541 |
+
"step": 380
|
| 542 |
+
},
|
| 543 |
+
{
|
| 544 |
+
"epoch": 0.7857142857142857,
|
| 545 |
+
"grad_norm": 239271.890625,
|
| 546 |
+
"learning_rate": 3.928571428571429e-06,
|
| 547 |
+
"loss": 1.3332,
|
| 548 |
+
"step": 385
|
| 549 |
+
},
|
| 550 |
+
{
|
| 551 |
+
"epoch": 0.7959183673469388,
|
| 552 |
+
"grad_norm": 167386.28125,
|
| 553 |
+
"learning_rate": 3.979591836734694e-06,
|
| 554 |
+
"loss": 1.429,
|
| 555 |
+
"step": 390
|
| 556 |
+
},
|
| 557 |
+
{
|
| 558 |
+
"epoch": 0.8061224489795918,
|
| 559 |
+
"grad_norm": 165701.9375,
|
| 560 |
+
"learning_rate": 4.03061224489796e-06,
|
| 561 |
+
"loss": 1.3167,
|
| 562 |
+
"step": 395
|
| 563 |
+
},
|
| 564 |
+
{
|
| 565 |
+
"epoch": 0.8163265306122449,
|
| 566 |
+
"grad_norm": 284793.46875,
|
| 567 |
+
"learning_rate": 4.081632653061225e-06,
|
| 568 |
+
"loss": 1.3867,
|
| 569 |
+
"step": 400
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.826530612244898,
|
| 573 |
+
"grad_norm": 159807.671875,
|
| 574 |
+
"learning_rate": 4.13265306122449e-06,
|
| 575 |
+
"loss": 1.404,
|
| 576 |
+
"step": 405
|
| 577 |
+
},
|
| 578 |
+
{
|
| 579 |
+
"epoch": 0.8367346938775511,
|
| 580 |
+
"grad_norm": 225698.859375,
|
| 581 |
+
"learning_rate": 4.183673469387755e-06,
|
| 582 |
+
"loss": 1.3842,
|
| 583 |
+
"step": 410
|
| 584 |
+
},
|
| 585 |
+
{
|
| 586 |
+
"epoch": 0.8469387755102041,
|
| 587 |
+
"grad_norm": 148945.90625,
|
| 588 |
+
"learning_rate": 4.234693877551021e-06,
|
| 589 |
+
"loss": 1.4605,
|
| 590 |
+
"step": 415
|
| 591 |
+
},
|
| 592 |
+
{
|
| 593 |
+
"epoch": 0.8571428571428571,
|
| 594 |
+
"grad_norm": 301350.15625,
|
| 595 |
+
"learning_rate": 4.2857142857142855e-06,
|
| 596 |
+
"loss": 1.4279,
|
| 597 |
+
"step": 420
|
| 598 |
+
},
|
| 599 |
+
{
|
| 600 |
+
"epoch": 0.8673469387755102,
|
| 601 |
+
"grad_norm": 216519.640625,
|
| 602 |
+
"learning_rate": 4.336734693877551e-06,
|
| 603 |
+
"loss": 1.3122,
|
| 604 |
+
"step": 425
|
| 605 |
+
},
|
| 606 |
+
{
|
| 607 |
+
"epoch": 0.8775510204081632,
|
| 608 |
+
"grad_norm": 243937.1875,
|
| 609 |
+
"learning_rate": 4.3877551020408165e-06,
|
| 610 |
+
"loss": 1.2084,
|
| 611 |
+
"step": 430
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.8877551020408163,
|
| 615 |
+
"grad_norm": 195260.390625,
|
| 616 |
+
"learning_rate": 4.438775510204082e-06,
|
| 617 |
+
"loss": 1.3024,
|
| 618 |
+
"step": 435
|
| 619 |
+
},
|
| 620 |
+
{
|
| 621 |
+
"epoch": 0.8979591836734694,
|
| 622 |
+
"grad_norm": 218123.65625,
|
| 623 |
+
"learning_rate": 4.489795918367348e-06,
|
| 624 |
+
"loss": 1.2303,
|
| 625 |
+
"step": 440
|
| 626 |
+
},
|
| 627 |
+
{
|
| 628 |
+
"epoch": 0.9081632653061225,
|
| 629 |
+
"grad_norm": 343990.03125,
|
| 630 |
+
"learning_rate": 4.540816326530613e-06,
|
| 631 |
+
"loss": 1.2012,
|
| 632 |
+
"step": 445
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.9183673469387755,
|
| 636 |
+
"grad_norm": 157196.390625,
|
| 637 |
+
"learning_rate": 4.591836734693878e-06,
|
| 638 |
+
"loss": 1.2632,
|
| 639 |
+
"step": 450
|
| 640 |
+
},
|
| 641 |
+
{
|
| 642 |
+
"epoch": 0.9285714285714286,
|
| 643 |
+
"grad_norm": 238103.46875,
|
| 644 |
+
"learning_rate": 4.642857142857144e-06,
|
| 645 |
+
"loss": 1.3131,
|
| 646 |
+
"step": 455
|
| 647 |
+
},
|
| 648 |
+
{
|
| 649 |
+
"epoch": 0.9387755102040817,
|
| 650 |
+
"grad_norm": 158442.625,
|
| 651 |
+
"learning_rate": 4.693877551020409e-06,
|
| 652 |
+
"loss": 1.2849,
|
| 653 |
+
"step": 460
|
| 654 |
+
},
|
| 655 |
+
{
|
| 656 |
+
"epoch": 0.9489795918367347,
|
| 657 |
+
"grad_norm": 178316.640625,
|
| 658 |
+
"learning_rate": 4.744897959183674e-06,
|
| 659 |
+
"loss": 1.0987,
|
| 660 |
+
"step": 465
|
| 661 |
+
},
|
| 662 |
+
{
|
| 663 |
+
"epoch": 0.9591836734693877,
|
| 664 |
+
"grad_norm": 240848.03125,
|
| 665 |
+
"learning_rate": 4.795918367346939e-06,
|
| 666 |
+
"loss": 1.154,
|
| 667 |
+
"step": 470
|
| 668 |
+
},
|
| 669 |
+
{
|
| 670 |
+
"epoch": 0.9693877551020408,
|
| 671 |
+
"grad_norm": 677730.8125,
|
| 672 |
+
"learning_rate": 4.846938775510204e-06,
|
| 673 |
+
"loss": 1.1682,
|
| 674 |
+
"step": 475
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.9795918367346939,
|
| 678 |
+
"grad_norm": 300622.875,
|
| 679 |
+
"learning_rate": 4.897959183673469e-06,
|
| 680 |
+
"loss": 1.1212,
|
| 681 |
+
"step": 480
|
| 682 |
+
},
|
| 683 |
+
{
|
| 684 |
+
"epoch": 0.9897959183673469,
|
| 685 |
+
"grad_norm": 234745.03125,
|
| 686 |
+
"learning_rate": 4.948979591836735e-06,
|
| 687 |
+
"loss": 1.1859,
|
| 688 |
+
"step": 485
|
| 689 |
+
},
|
| 690 |
+
{
|
| 691 |
+
"epoch": 1.0,
|
| 692 |
+
"grad_norm": 254966.984375,
|
| 693 |
+
"learning_rate": 5e-06,
|
| 694 |
+
"loss": 1.1763,
|
| 695 |
+
"step": 490
|
| 696 |
+
},
|
| 697 |
+
{
|
| 698 |
+
"epoch": 1.0,
|
| 699 |
+
"eval_accuracy": 0.8883299798792756,
|
| 700 |
+
"eval_loss": 0.968666136264801,
|
| 701 |
+
"eval_runtime": 208.5533,
|
| 702 |
+
"eval_samples_per_second": 14.299,
|
| 703 |
+
"eval_steps_per_second": 0.599,
|
| 704 |
+
"step": 490
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 1.010204081632653,
|
| 708 |
+
"grad_norm": 227501.953125,
|
| 709 |
+
"learning_rate": 4.994331065759638e-06,
|
| 710 |
+
"loss": 1.1466,
|
| 711 |
+
"step": 495
|
| 712 |
+
},
|
| 713 |
+
{
|
| 714 |
+
"epoch": 1.0204081632653061,
|
| 715 |
+
"grad_norm": 184310.828125,
|
| 716 |
+
"learning_rate": 4.9886621315192745e-06,
|
| 717 |
+
"loss": 1.0584,
|
| 718 |
+
"step": 500
|
| 719 |
+
},
|
| 720 |
+
{
|
| 721 |
+
"epoch": 1.030612244897959,
|
| 722 |
+
"grad_norm": 568293.25,
|
| 723 |
+
"learning_rate": 4.982993197278912e-06,
|
| 724 |
+
"loss": 1.0574,
|
| 725 |
+
"step": 505
|
| 726 |
+
},
|
| 727 |
+
{
|
| 728 |
+
"epoch": 1.0408163265306123,
|
| 729 |
+
"grad_norm": 208197.4375,
|
| 730 |
+
"learning_rate": 4.9773242630385495e-06,
|
| 731 |
+
"loss": 1.0964,
|
| 732 |
+
"step": 510
|
| 733 |
+
},
|
| 734 |
+
{
|
| 735 |
+
"epoch": 1.0510204081632653,
|
| 736 |
+
"grad_norm": 302428.34375,
|
| 737 |
+
"learning_rate": 4.971655328798186e-06,
|
| 738 |
+
"loss": 1.1017,
|
| 739 |
+
"step": 515
|
| 740 |
+
},
|
| 741 |
+
{
|
| 742 |
+
"epoch": 1.0612244897959184,
|
| 743 |
+
"grad_norm": 331227.1875,
|
| 744 |
+
"learning_rate": 4.965986394557824e-06,
|
| 745 |
+
"loss": 1.0954,
|
| 746 |
+
"step": 520
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 1.0714285714285714,
|
| 750 |
+
"grad_norm": 525813.5625,
|
| 751 |
+
"learning_rate": 4.960317460317461e-06,
|
| 752 |
+
"loss": 1.061,
|
| 753 |
+
"step": 525
|
| 754 |
+
},
|
| 755 |
+
{
|
| 756 |
+
"epoch": 1.0816326530612246,
|
| 757 |
+
"grad_norm": 578634.125,
|
| 758 |
+
"learning_rate": 4.954648526077098e-06,
|
| 759 |
+
"loss": 1.135,
|
| 760 |
+
"step": 530
|
| 761 |
+
},
|
| 762 |
+
{
|
| 763 |
+
"epoch": 1.0918367346938775,
|
| 764 |
+
"grad_norm": 208531.90625,
|
| 765 |
+
"learning_rate": 4.948979591836735e-06,
|
| 766 |
+
"loss": 0.9104,
|
| 767 |
+
"step": 535
|
| 768 |
+
},
|
| 769 |
+
{
|
| 770 |
+
"epoch": 1.1020408163265305,
|
| 771 |
+
"grad_norm": 621899.75,
|
| 772 |
+
"learning_rate": 4.943310657596373e-06,
|
| 773 |
+
"loss": 1.0909,
|
| 774 |
+
"step": 540
|
| 775 |
+
},
|
| 776 |
+
{
|
| 777 |
+
"epoch": 1.1122448979591837,
|
| 778 |
+
"grad_norm": 362684.3125,
|
| 779 |
+
"learning_rate": 4.9376417233560094e-06,
|
| 780 |
+
"loss": 0.8759,
|
| 781 |
+
"step": 545
|
| 782 |
+
},
|
| 783 |
+
{
|
| 784 |
+
"epoch": 1.1224489795918366,
|
| 785 |
+
"grad_norm": 274587.8125,
|
| 786 |
+
"learning_rate": 4.931972789115647e-06,
|
| 787 |
+
"loss": 0.9977,
|
| 788 |
+
"step": 550
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 1.1326530612244898,
|
| 792 |
+
"grad_norm": 379287.3125,
|
| 793 |
+
"learning_rate": 4.9263038548752836e-06,
|
| 794 |
+
"loss": 1.013,
|
| 795 |
+
"step": 555
|
| 796 |
+
},
|
| 797 |
+
{
|
| 798 |
+
"epoch": 1.1428571428571428,
|
| 799 |
+
"grad_norm": 277623.0,
|
| 800 |
+
"learning_rate": 4.920634920634921e-06,
|
| 801 |
+
"loss": 0.9425,
|
| 802 |
+
"step": 560
|
| 803 |
+
},
|
| 804 |
+
{
|
| 805 |
+
"epoch": 1.153061224489796,
|
| 806 |
+
"grad_norm": 341604.53125,
|
| 807 |
+
"learning_rate": 4.914965986394558e-06,
|
| 808 |
+
"loss": 0.9697,
|
| 809 |
+
"step": 565
|
| 810 |
+
},
|
| 811 |
+
{
|
| 812 |
+
"epoch": 1.163265306122449,
|
| 813 |
+
"grad_norm": 742070.625,
|
| 814 |
+
"learning_rate": 4.909297052154195e-06,
|
| 815 |
+
"loss": 1.1181,
|
| 816 |
+
"step": 570
|
| 817 |
+
},
|
| 818 |
+
{
|
| 819 |
+
"epoch": 1.1734693877551021,
|
| 820 |
+
"grad_norm": 225296.84375,
|
| 821 |
+
"learning_rate": 4.903628117913833e-06,
|
| 822 |
+
"loss": 0.9465,
|
| 823 |
+
"step": 575
|
| 824 |
+
},
|
| 825 |
+
{
|
| 826 |
+
"epoch": 1.183673469387755,
|
| 827 |
+
"grad_norm": 952885.3125,
|
| 828 |
+
"learning_rate": 4.897959183673469e-06,
|
| 829 |
+
"loss": 1.1135,
|
| 830 |
+
"step": 580
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 1.193877551020408,
|
| 834 |
+
"grad_norm": 527028.875,
|
| 835 |
+
"learning_rate": 4.892290249433107e-06,
|
| 836 |
+
"loss": 1.109,
|
| 837 |
+
"step": 585
|
| 838 |
+
},
|
| 839 |
+
{
|
| 840 |
+
"epoch": 1.2040816326530612,
|
| 841 |
+
"grad_norm": 243262.0625,
|
| 842 |
+
"learning_rate": 4.886621315192744e-06,
|
| 843 |
+
"loss": 0.9313,
|
| 844 |
+
"step": 590
|
| 845 |
+
},
|
| 846 |
+
{
|
| 847 |
+
"epoch": 1.2142857142857142,
|
| 848 |
+
"grad_norm": 287492.28125,
|
| 849 |
+
"learning_rate": 4.880952380952381e-06,
|
| 850 |
+
"loss": 0.9603,
|
| 851 |
+
"step": 595
|
| 852 |
+
},
|
| 853 |
+
{
|
| 854 |
+
"epoch": 1.2244897959183674,
|
| 855 |
+
"grad_norm": 400522.90625,
|
| 856 |
+
"learning_rate": 4.8752834467120185e-06,
|
| 857 |
+
"loss": 0.9105,
|
| 858 |
+
"step": 600
|
| 859 |
+
},
|
| 860 |
+
{
|
| 861 |
+
"epoch": 1.2346938775510203,
|
| 862 |
+
"grad_norm": 190966.40625,
|
| 863 |
+
"learning_rate": 4.869614512471656e-06,
|
| 864 |
+
"loss": 0.9654,
|
| 865 |
+
"step": 605
|
| 866 |
+
},
|
| 867 |
+
{
|
| 868 |
+
"epoch": 1.2448979591836735,
|
| 869 |
+
"grad_norm": 223976.515625,
|
| 870 |
+
"learning_rate": 4.863945578231293e-06,
|
| 871 |
+
"loss": 0.9511,
|
| 872 |
+
"step": 610
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 1.2551020408163265,
|
| 876 |
+
"grad_norm": 345172.09375,
|
| 877 |
+
"learning_rate": 4.85827664399093e-06,
|
| 878 |
+
"loss": 0.8092,
|
| 879 |
+
"step": 615
|
| 880 |
+
},
|
| 881 |
+
{
|
| 882 |
+
"epoch": 1.2653061224489797,
|
| 883 |
+
"grad_norm": 378395.75,
|
| 884 |
+
"learning_rate": 4.852607709750568e-06,
|
| 885 |
+
"loss": 0.9777,
|
| 886 |
+
"step": 620
|
| 887 |
+
},
|
| 888 |
+
{
|
| 889 |
+
"epoch": 1.2755102040816326,
|
| 890 |
+
"grad_norm": 408057.8125,
|
| 891 |
+
"learning_rate": 4.846938775510204e-06,
|
| 892 |
+
"loss": 1.006,
|
| 893 |
+
"step": 625
|
| 894 |
+
},
|
| 895 |
+
{
|
| 896 |
+
"epoch": 1.2857142857142856,
|
| 897 |
+
"grad_norm": 254345.203125,
|
| 898 |
+
"learning_rate": 4.841269841269842e-06,
|
| 899 |
+
"loss": 0.9107,
|
| 900 |
+
"step": 630
|
| 901 |
+
},
|
| 902 |
+
{
|
| 903 |
+
"epoch": 1.2959183673469388,
|
| 904 |
+
"grad_norm": 465929.21875,
|
| 905 |
+
"learning_rate": 4.835600907029479e-06,
|
| 906 |
+
"loss": 1.0237,
|
| 907 |
+
"step": 635
|
| 908 |
+
},
|
| 909 |
+
{
|
| 910 |
+
"epoch": 1.306122448979592,
|
| 911 |
+
"grad_norm": 665911.5625,
|
| 912 |
+
"learning_rate": 4.829931972789116e-06,
|
| 913 |
+
"loss": 0.886,
|
| 914 |
+
"step": 640
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.316326530612245,
|
| 918 |
+
"grad_norm": 541784.25,
|
| 919 |
+
"learning_rate": 4.824263038548753e-06,
|
| 920 |
+
"loss": 0.9266,
|
| 921 |
+
"step": 645
|
| 922 |
+
},
|
| 923 |
+
{
|
| 924 |
+
"epoch": 1.3265306122448979,
|
| 925 |
+
"grad_norm": 181121.25,
|
| 926 |
+
"learning_rate": 4.818594104308391e-06,
|
| 927 |
+
"loss": 0.9289,
|
| 928 |
+
"step": 650
|
| 929 |
+
},
|
| 930 |
+
{
|
| 931 |
+
"epoch": 1.336734693877551,
|
| 932 |
+
"grad_norm": 465088.90625,
|
| 933 |
+
"learning_rate": 4.8129251700680275e-06,
|
| 934 |
+
"loss": 0.8976,
|
| 935 |
+
"step": 655
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 1.346938775510204,
|
| 939 |
+
"grad_norm": 309529.28125,
|
| 940 |
+
"learning_rate": 4.807256235827665e-06,
|
| 941 |
+
"loss": 0.9182,
|
| 942 |
+
"step": 660
|
| 943 |
+
},
|
| 944 |
+
{
|
| 945 |
+
"epoch": 1.3571428571428572,
|
| 946 |
+
"grad_norm": 672721.125,
|
| 947 |
+
"learning_rate": 4.8015873015873025e-06,
|
| 948 |
+
"loss": 0.8919,
|
| 949 |
+
"step": 665
|
| 950 |
+
},
|
| 951 |
+
{
|
| 952 |
+
"epoch": 1.3673469387755102,
|
| 953 |
+
"grad_norm": 300983.96875,
|
| 954 |
+
"learning_rate": 4.795918367346939e-06,
|
| 955 |
+
"loss": 0.7491,
|
| 956 |
+
"step": 670
|
| 957 |
+
},
|
| 958 |
+
{
|
| 959 |
+
"epoch": 1.3775510204081631,
|
| 960 |
+
"grad_norm": 168024.203125,
|
| 961 |
+
"learning_rate": 4.790249433106577e-06,
|
| 962 |
+
"loss": 0.803,
|
| 963 |
+
"step": 675
|
| 964 |
+
},
|
| 965 |
+
{
|
| 966 |
+
"epoch": 1.3877551020408163,
|
| 967 |
+
"grad_norm": 252603.5,
|
| 968 |
+
"learning_rate": 4.784580498866213e-06,
|
| 969 |
+
"loss": 0.8481,
|
| 970 |
+
"step": 680
|
| 971 |
+
},
|
| 972 |
+
{
|
| 973 |
+
"epoch": 1.3979591836734695,
|
| 974 |
+
"grad_norm": 495847.71875,
|
| 975 |
+
"learning_rate": 4.778911564625851e-06,
|
| 976 |
+
"loss": 0.9796,
|
| 977 |
+
"step": 685
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 1.4081632653061225,
|
| 981 |
+
"grad_norm": 213929.046875,
|
| 982 |
+
"learning_rate": 4.773242630385488e-06,
|
| 983 |
+
"loss": 0.8509,
|
| 984 |
+
"step": 690
|
| 985 |
+
},
|
| 986 |
+
{
|
| 987 |
+
"epoch": 1.4183673469387754,
|
| 988 |
+
"grad_norm": 203737.90625,
|
| 989 |
+
"learning_rate": 4.767573696145125e-06,
|
| 990 |
+
"loss": 0.8133,
|
| 991 |
+
"step": 695
|
| 992 |
+
},
|
| 993 |
+
{
|
| 994 |
+
"epoch": 1.4285714285714286,
|
| 995 |
+
"grad_norm": 340089.6875,
|
| 996 |
+
"learning_rate": 4.761904761904762e-06,
|
| 997 |
+
"loss": 0.8628,
|
| 998 |
+
"step": 700
|
| 999 |
+
},
|
| 1000 |
+
{
|
| 1001 |
+
"epoch": 1.4387755102040816,
|
| 1002 |
+
"grad_norm": 292163.90625,
|
| 1003 |
+
"learning_rate": 4.756235827664399e-06,
|
| 1004 |
+
"loss": 0.9341,
|
| 1005 |
+
"step": 705
|
| 1006 |
+
},
|
| 1007 |
+
{
|
| 1008 |
+
"epoch": 1.4489795918367347,
|
| 1009 |
+
"grad_norm": 266707.96875,
|
| 1010 |
+
"learning_rate": 4.7505668934240365e-06,
|
| 1011 |
+
"loss": 1.0158,
|
| 1012 |
+
"step": 710
|
| 1013 |
+
},
|
| 1014 |
+
{
|
| 1015 |
+
"epoch": 1.4591836734693877,
|
| 1016 |
+
"grad_norm": 389876.40625,
|
| 1017 |
+
"learning_rate": 4.744897959183674e-06,
|
| 1018 |
+
"loss": 0.9575,
|
| 1019 |
+
"step": 715
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 1.469387755102041,
|
| 1023 |
+
"grad_norm": 699534.0625,
|
| 1024 |
+
"learning_rate": 4.739229024943311e-06,
|
| 1025 |
+
"loss": 0.9293,
|
| 1026 |
+
"step": 720
|
| 1027 |
+
},
|
| 1028 |
+
{
|
| 1029 |
+
"epoch": 1.4795918367346939,
|
| 1030 |
+
"grad_norm": 495432.5,
|
| 1031 |
+
"learning_rate": 4.733560090702948e-06,
|
| 1032 |
+
"loss": 0.7233,
|
| 1033 |
+
"step": 725
|
| 1034 |
+
},
|
| 1035 |
+
{
|
| 1036 |
+
"epoch": 1.489795918367347,
|
| 1037 |
+
"grad_norm": 297580.71875,
|
| 1038 |
+
"learning_rate": 4.727891156462586e-06,
|
| 1039 |
+
"loss": 0.7624,
|
| 1040 |
+
"step": 730
|
| 1041 |
+
},
|
| 1042 |
+
{
|
| 1043 |
+
"epoch": 1.5,
|
| 1044 |
+
"grad_norm": 214458.125,
|
| 1045 |
+
"learning_rate": 4.722222222222222e-06,
|
| 1046 |
+
"loss": 0.8704,
|
| 1047 |
+
"step": 735
|
| 1048 |
+
},
|
| 1049 |
+
{
|
| 1050 |
+
"epoch": 1.510204081632653,
|
| 1051 |
+
"grad_norm": 483185.90625,
|
| 1052 |
+
"learning_rate": 4.71655328798186e-06,
|
| 1053 |
+
"loss": 0.7841,
|
| 1054 |
+
"step": 740
|
| 1055 |
+
},
|
| 1056 |
+
{
|
| 1057 |
+
"epoch": 1.5204081632653061,
|
| 1058 |
+
"grad_norm": 279528.75,
|
| 1059 |
+
"learning_rate": 4.710884353741497e-06,
|
| 1060 |
+
"loss": 0.8634,
|
| 1061 |
+
"step": 745
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 1.5306122448979593,
|
| 1065 |
+
"grad_norm": 398520.15625,
|
| 1066 |
+
"learning_rate": 4.705215419501134e-06,
|
| 1067 |
+
"loss": 0.902,
|
| 1068 |
+
"step": 750
|
| 1069 |
+
},
|
| 1070 |
+
{
|
| 1071 |
+
"epoch": 1.5408163265306123,
|
| 1072 |
+
"grad_norm": 343986.875,
|
| 1073 |
+
"learning_rate": 4.699546485260771e-06,
|
| 1074 |
+
"loss": 0.8896,
|
| 1075 |
+
"step": 755
|
| 1076 |
+
},
|
| 1077 |
+
{
|
| 1078 |
+
"epoch": 1.5510204081632653,
|
| 1079 |
+
"grad_norm": 941463.0625,
|
| 1080 |
+
"learning_rate": 4.693877551020409e-06,
|
| 1081 |
+
"loss": 0.8483,
|
| 1082 |
+
"step": 760
|
| 1083 |
+
},
|
| 1084 |
+
{
|
| 1085 |
+
"epoch": 1.5612244897959182,
|
| 1086 |
+
"grad_norm": 382110.125,
|
| 1087 |
+
"learning_rate": 4.6882086167800455e-06,
|
| 1088 |
+
"loss": 0.944,
|
| 1089 |
+
"step": 765
|
| 1090 |
+
},
|
| 1091 |
+
{
|
| 1092 |
+
"epoch": 1.5714285714285714,
|
| 1093 |
+
"grad_norm": 321413.15625,
|
| 1094 |
+
"learning_rate": 4.682539682539683e-06,
|
| 1095 |
+
"loss": 0.8019,
|
| 1096 |
+
"step": 770
|
| 1097 |
+
},
|
| 1098 |
+
{
|
| 1099 |
+
"epoch": 1.5816326530612246,
|
| 1100 |
+
"grad_norm": 404689.875,
|
| 1101 |
+
"learning_rate": 4.6768707482993205e-06,
|
| 1102 |
+
"loss": 0.7069,
|
| 1103 |
+
"step": 775
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 1.5918367346938775,
|
| 1107 |
+
"grad_norm": 653981.25,
|
| 1108 |
+
"learning_rate": 4.671201814058957e-06,
|
| 1109 |
+
"loss": 0.9424,
|
| 1110 |
+
"step": 780
|
| 1111 |
+
},
|
| 1112 |
+
{
|
| 1113 |
+
"epoch": 1.6020408163265305,
|
| 1114 |
+
"grad_norm": 387232.34375,
|
| 1115 |
+
"learning_rate": 4.665532879818595e-06,
|
| 1116 |
+
"loss": 0.9593,
|
| 1117 |
+
"step": 785
|
| 1118 |
+
},
|
| 1119 |
+
{
|
| 1120 |
+
"epoch": 1.6122448979591837,
|
| 1121 |
+
"grad_norm": 527808.625,
|
| 1122 |
+
"learning_rate": 4.659863945578232e-06,
|
| 1123 |
+
"loss": 0.8147,
|
| 1124 |
+
"step": 790
|
| 1125 |
+
},
|
| 1126 |
+
{
|
| 1127 |
+
"epoch": 1.6224489795918369,
|
| 1128 |
+
"grad_norm": 350020.8125,
|
| 1129 |
+
"learning_rate": 4.654195011337869e-06,
|
| 1130 |
+
"loss": 1.0195,
|
| 1131 |
+
"step": 795
|
| 1132 |
+
},
|
| 1133 |
+
{
|
| 1134 |
+
"epoch": 1.6326530612244898,
|
| 1135 |
+
"grad_norm": 510409.9375,
|
| 1136 |
+
"learning_rate": 4.648526077097506e-06,
|
| 1137 |
+
"loss": 0.865,
|
| 1138 |
+
"step": 800
|
| 1139 |
+
},
|
| 1140 |
+
{
|
| 1141 |
+
"epoch": 1.6428571428571428,
|
| 1142 |
+
"grad_norm": 238838.65625,
|
| 1143 |
+
"learning_rate": 4.642857142857144e-06,
|
| 1144 |
+
"loss": 0.8115,
|
| 1145 |
+
"step": 805
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 1.6530612244897958,
|
| 1149 |
+
"grad_norm": 266616.84375,
|
| 1150 |
+
"learning_rate": 4.6371882086167804e-06,
|
| 1151 |
+
"loss": 0.8091,
|
| 1152 |
+
"step": 810
|
| 1153 |
+
},
|
| 1154 |
+
{
|
| 1155 |
+
"epoch": 1.663265306122449,
|
| 1156 |
+
"grad_norm": 398578.84375,
|
| 1157 |
+
"learning_rate": 4.631519274376418e-06,
|
| 1158 |
+
"loss": 0.6485,
|
| 1159 |
+
"step": 815
|
| 1160 |
+
},
|
| 1161 |
+
{
|
| 1162 |
+
"epoch": 1.6734693877551021,
|
| 1163 |
+
"grad_norm": 242011.765625,
|
| 1164 |
+
"learning_rate": 4.6258503401360546e-06,
|
| 1165 |
+
"loss": 0.8242,
|
| 1166 |
+
"step": 820
|
| 1167 |
+
},
|
| 1168 |
+
{
|
| 1169 |
+
"epoch": 1.683673469387755,
|
| 1170 |
+
"grad_norm": 646366.6875,
|
| 1171 |
+
"learning_rate": 4.620181405895692e-06,
|
| 1172 |
+
"loss": 0.7616,
|
| 1173 |
+
"step": 825
|
| 1174 |
+
},
|
| 1175 |
+
{
|
| 1176 |
+
"epoch": 1.693877551020408,
|
| 1177 |
+
"grad_norm": 447861.9375,
|
| 1178 |
+
"learning_rate": 4.614512471655329e-06,
|
| 1179 |
+
"loss": 0.8694,
|
| 1180 |
+
"step": 830
|
| 1181 |
+
},
|
| 1182 |
+
{
|
| 1183 |
+
"epoch": 1.7040816326530612,
|
| 1184 |
+
"grad_norm": 422529.3125,
|
| 1185 |
+
"learning_rate": 4.608843537414966e-06,
|
| 1186 |
+
"loss": 0.9344,
|
| 1187 |
+
"step": 835
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 1.7142857142857144,
|
| 1191 |
+
"grad_norm": 448338.375,
|
| 1192 |
+
"learning_rate": 4.603174603174604e-06,
|
| 1193 |
+
"loss": 0.8459,
|
| 1194 |
+
"step": 840
|
| 1195 |
+
},
|
| 1196 |
+
{
|
| 1197 |
+
"epoch": 1.7244897959183674,
|
| 1198 |
+
"grad_norm": 528423.75,
|
| 1199 |
+
"learning_rate": 4.59750566893424e-06,
|
| 1200 |
+
"loss": 0.8312,
|
| 1201 |
+
"step": 845
|
| 1202 |
+
},
|
| 1203 |
+
{
|
| 1204 |
+
"epoch": 1.7346938775510203,
|
| 1205 |
+
"grad_norm": 422879.84375,
|
| 1206 |
+
"learning_rate": 4.591836734693878e-06,
|
| 1207 |
+
"loss": 0.8151,
|
| 1208 |
+
"step": 850
|
| 1209 |
+
},
|
| 1210 |
+
{
|
| 1211 |
+
"epoch": 1.7448979591836735,
|
| 1212 |
+
"grad_norm": 543066.5,
|
| 1213 |
+
"learning_rate": 4.5861678004535145e-06,
|
| 1214 |
+
"loss": 0.8013,
|
| 1215 |
+
"step": 855
|
| 1216 |
+
},
|
| 1217 |
+
{
|
| 1218 |
+
"epoch": 1.7551020408163265,
|
| 1219 |
+
"grad_norm": 362556.6875,
|
| 1220 |
+
"learning_rate": 4.580498866213152e-06,
|
| 1221 |
+
"loss": 0.9636,
|
| 1222 |
+
"step": 860
|
| 1223 |
+
},
|
| 1224 |
+
{
|
| 1225 |
+
"epoch": 1.7653061224489797,
|
| 1226 |
+
"grad_norm": 305676.09375,
|
| 1227 |
+
"learning_rate": 4.5748299319727895e-06,
|
| 1228 |
+
"loss": 0.8208,
|
| 1229 |
+
"step": 865
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 1.7755102040816326,
|
| 1233 |
+
"grad_norm": 477851.96875,
|
| 1234 |
+
"learning_rate": 4.569160997732427e-06,
|
| 1235 |
+
"loss": 0.8109,
|
| 1236 |
+
"step": 870
|
| 1237 |
+
},
|
| 1238 |
+
{
|
| 1239 |
+
"epoch": 1.7857142857142856,
|
| 1240 |
+
"grad_norm": 380819.1875,
|
| 1241 |
+
"learning_rate": 4.563492063492064e-06,
|
| 1242 |
+
"loss": 0.8019,
|
| 1243 |
+
"step": 875
|
| 1244 |
+
},
|
| 1245 |
+
{
|
| 1246 |
+
"epoch": 1.7959183673469388,
|
| 1247 |
+
"grad_norm": 276121.21875,
|
| 1248 |
+
"learning_rate": 4.557823129251701e-06,
|
| 1249 |
+
"loss": 0.7365,
|
| 1250 |
+
"step": 880
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 1.806122448979592,
|
| 1254 |
+
"grad_norm": 807004.5625,
|
| 1255 |
+
"learning_rate": 4.5521541950113386e-06,
|
| 1256 |
+
"loss": 0.8284,
|
| 1257 |
+
"step": 885
|
| 1258 |
+
},
|
| 1259 |
+
{
|
| 1260 |
+
"epoch": 1.816326530612245,
|
| 1261 |
+
"grad_norm": 612872.5625,
|
| 1262 |
+
"learning_rate": 4.546485260770975e-06,
|
| 1263 |
+
"loss": 0.9153,
|
| 1264 |
+
"step": 890
|
| 1265 |
+
},
|
| 1266 |
+
{
|
| 1267 |
+
"epoch": 1.8265306122448979,
|
| 1268 |
+
"grad_norm": 384264.40625,
|
| 1269 |
+
"learning_rate": 4.540816326530613e-06,
|
| 1270 |
+
"loss": 0.8697,
|
| 1271 |
+
"step": 895
|
| 1272 |
+
},
|
| 1273 |
+
{
|
| 1274 |
+
"epoch": 1.836734693877551,
|
| 1275 |
+
"grad_norm": 323550.15625,
|
| 1276 |
+
"learning_rate": 4.53514739229025e-06,
|
| 1277 |
+
"loss": 0.816,
|
| 1278 |
+
"step": 900
|
| 1279 |
+
},
|
| 1280 |
+
{
|
| 1281 |
+
"epoch": 1.8469387755102042,
|
| 1282 |
+
"grad_norm": 588055.625,
|
| 1283 |
+
"learning_rate": 4.529478458049887e-06,
|
| 1284 |
+
"loss": 0.9765,
|
| 1285 |
+
"step": 905
|
| 1286 |
+
},
|
| 1287 |
+
{
|
| 1288 |
+
"epoch": 1.8571428571428572,
|
| 1289 |
+
"grad_norm": 1386326.375,
|
| 1290 |
+
"learning_rate": 4.523809523809524e-06,
|
| 1291 |
+
"loss": 0.8051,
|
| 1292 |
+
"step": 910
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 1.8673469387755102,
|
| 1296 |
+
"grad_norm": 351753.28125,
|
| 1297 |
+
"learning_rate": 4.518140589569162e-06,
|
| 1298 |
+
"loss": 0.8377,
|
| 1299 |
+
"step": 915
|
| 1300 |
+
},
|
| 1301 |
+
{
|
| 1302 |
+
"epoch": 1.8775510204081631,
|
| 1303 |
+
"grad_norm": 547828.25,
|
| 1304 |
+
"learning_rate": 4.5124716553287985e-06,
|
| 1305 |
+
"loss": 0.6776,
|
| 1306 |
+
"step": 920
|
| 1307 |
+
},
|
| 1308 |
+
{
|
| 1309 |
+
"epoch": 1.8877551020408163,
|
| 1310 |
+
"grad_norm": 332551.40625,
|
| 1311 |
+
"learning_rate": 4.506802721088436e-06,
|
| 1312 |
+
"loss": 0.9434,
|
| 1313 |
+
"step": 925
|
| 1314 |
+
},
|
| 1315 |
+
{
|
| 1316 |
+
"epoch": 1.8979591836734695,
|
| 1317 |
+
"grad_norm": 597800.125,
|
| 1318 |
+
"learning_rate": 4.5011337868480735e-06,
|
| 1319 |
+
"loss": 0.6165,
|
| 1320 |
+
"step": 930
|
| 1321 |
+
},
|
| 1322 |
+
{
|
| 1323 |
+
"epoch": 1.9081632653061225,
|
| 1324 |
+
"grad_norm": 598426.1875,
|
| 1325 |
+
"learning_rate": 4.49546485260771e-06,
|
| 1326 |
+
"loss": 0.9668,
|
| 1327 |
+
"step": 935
|
| 1328 |
+
},
|
| 1329 |
+
{
|
| 1330 |
+
"epoch": 1.9183673469387754,
|
| 1331 |
+
"grad_norm": 508502.5,
|
| 1332 |
+
"learning_rate": 4.489795918367348e-06,
|
| 1333 |
+
"loss": 0.8707,
|
| 1334 |
+
"step": 940
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 1.9285714285714286,
|
| 1338 |
+
"grad_norm": 825221.0625,
|
| 1339 |
+
"learning_rate": 4.484126984126984e-06,
|
| 1340 |
+
"loss": 0.8437,
|
| 1341 |
+
"step": 945
|
| 1342 |
+
},
|
| 1343 |
+
{
|
| 1344 |
+
"epoch": 1.9387755102040818,
|
| 1345 |
+
"grad_norm": 674555.8125,
|
| 1346 |
+
"learning_rate": 4.478458049886622e-06,
|
| 1347 |
+
"loss": 0.7054,
|
| 1348 |
+
"step": 950
|
| 1349 |
+
},
|
| 1350 |
+
{
|
| 1351 |
+
"epoch": 1.9489795918367347,
|
| 1352 |
+
"grad_norm": 547481.0,
|
| 1353 |
+
"learning_rate": 4.472789115646259e-06,
|
| 1354 |
+
"loss": 0.7896,
|
| 1355 |
+
"step": 955
|
| 1356 |
+
},
|
| 1357 |
+
{
|
| 1358 |
+
"epoch": 1.9591836734693877,
|
| 1359 |
+
"grad_norm": 869727.6875,
|
| 1360 |
+
"learning_rate": 4.467120181405896e-06,
|
| 1361 |
+
"loss": 0.8161,
|
| 1362 |
+
"step": 960
|
| 1363 |
+
},
|
| 1364 |
+
{
|
| 1365 |
+
"epoch": 1.9693877551020407,
|
| 1366 |
+
"grad_norm": 517315.71875,
|
| 1367 |
+
"learning_rate": 4.461451247165533e-06,
|
| 1368 |
+
"loss": 0.8131,
|
| 1369 |
+
"step": 965
|
| 1370 |
+
},
|
| 1371 |
+
{
|
| 1372 |
+
"epoch": 1.9795918367346939,
|
| 1373 |
+
"grad_norm": 651501.0625,
|
| 1374 |
+
"learning_rate": 4.45578231292517e-06,
|
| 1375 |
+
"loss": 0.6217,
|
| 1376 |
+
"step": 970
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 1.989795918367347,
|
| 1380 |
+
"grad_norm": 559789.5625,
|
| 1381 |
+
"learning_rate": 4.4501133786848075e-06,
|
| 1382 |
+
"loss": 0.7849,
|
| 1383 |
+
"step": 975
|
| 1384 |
+
},
|
| 1385 |
+
{
|
| 1386 |
+
"epoch": 2.0,
|
| 1387 |
+
"grad_norm": 586872.375,
|
| 1388 |
+
"learning_rate": 4.444444444444444e-06,
|
| 1389 |
+
"loss": 0.7778,
|
| 1390 |
+
"step": 980
|
| 1391 |
+
},
|
| 1392 |
+
{
|
| 1393 |
+
"epoch": 2.0,
|
| 1394 |
+
"eval_accuracy": 0.8779342723004695,
|
| 1395 |
+
"eval_loss": 0.5649504661560059,
|
| 1396 |
+
"eval_runtime": 182.7428,
|
| 1397 |
+
"eval_samples_per_second": 16.318,
|
| 1398 |
+
"eval_steps_per_second": 0.684,
|
| 1399 |
+
"step": 980
|
| 1400 |
+
}
|
| 1401 |
+
],
|
| 1402 |
+
"logging_steps": 5,
|
| 1403 |
+
"max_steps": 4900,
|
| 1404 |
+
"num_input_tokens_seen": 0,
|
| 1405 |
+
"num_train_epochs": 10,
|
| 1406 |
+
"save_steps": 500,
|
| 1407 |
+
"stateful_callbacks": {
|
| 1408 |
+
"TrainerControl": {
|
| 1409 |
+
"args": {
|
| 1410 |
+
"should_epoch_stop": false,
|
| 1411 |
+
"should_evaluate": false,
|
| 1412 |
+
"should_log": false,
|
| 1413 |
+
"should_save": true,
|
| 1414 |
+
"should_training_stop": false
|
| 1415 |
+
},
|
| 1416 |
+
"attributes": {}
|
| 1417 |
+
}
|
| 1418 |
+
},
|
| 1419 |
+
"total_flos": 2.99358138673152e+18,
|
| 1420 |
+
"train_batch_size": 24,
|
| 1421 |
+
"trial_name": null,
|
| 1422 |
+
"trial_params": null
|
| 1423 |
+
}
|
checkpoint-980/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:65e022e2b943d9a83eff2469e42d3ed5d9859bd381fb675a108709af3c02f25f
|
| 3 |
+
size 5176
|